EP0056600A2

EP0056600A2 - Modular structure of PCM-switched distributed control and distributed-diagnostic network

Info

Publication number: EP0056600A2
Application number: EP82100149A
Authority: EP
Inventors: Piero Belforte; Mario Bondonno; Enzo Garetti; Giancarlo Guaschino; Luciano Pilati
Original assignee: CSELT Centro Studi e Laboratori Telecomunicazioni SpA
Current assignee: Telecom Italia SpA
Priority date: 1981-01-15
Filing date: 1982-01-11
Publication date: 1982-07-28
Also published as: AU545679B2; ATE10564T1; JPH0449317B2; DK13482A; BR8200049A; US4473900A; DE3261306D1; CA1195760A; DK163631C; ES8303855A1; AU7926182A; DK163631B; ES508539A0; IT8167036A0; IT1143268B; EP0056600B1; JPS57138292A; DE56600T1; EP0056600A3

Abstract

Interconnection unit equipped with a microprocessor- controlled unit and auxiliary circuits specialized in diagnostic, allowing the functions of diagnostic, trouble localization and reconfiguration to be carried out by the single building blocks, without requiring the generation of artificial test traffic.

Description

The present invention concerns PCM-switched systems and more particularly it concerns the embodiment of modularly expandable, distributed-control, PCM-switched networks, which also comprise, control and diagnostic functions, locally implemented by commercially available microprocessors.
The main properties an up-to-date switching network ought to have are the following:

a) capacity ranging from few hundred channels (local exchanges) to hundreds of-thousands of channels (transit exchanges), by using a small number of different kinds of switching elementary modules, i. e. minimum building units easy to replace;
b) use of high-flexibility switching elementary modules so that they may cover as wide a range as possible of applications in the switching field, without significantly suffering from technological component development;
c) expanding network capacity in a modular way, by maintaining a good efficiency in the ratio between used hardware and number of equipped channels;
d) network control structure independent of the centralized telephone control of the exchange, so as to make it free from the network management;
e) network control structure, if possible of the distributed type, so that the ratio between the installedproces sing power and the number of channels to be served, may be optimized;
f) network diagnostic independent of the whole diagnostic function and decentralized to the level of single switching elementary module so that a quick failure detection and localization may be obtained;
g) network blocking characteristics capable of assuring negligible losses in whatever chosen structure and under whatever equipment condition for maximum traffic conditions envisaged;
h) network control with minimal connection actuating times and capacity of carrying out as many connecting requests as possible;
i) minimal transit delay of the PCM samples through the network;
1) minimum encumbrance and power dissipation in comparison with the number of equipped channels;
m) characteristic of network degradation, that in case of failure, drastically limits the number of channels gone out of service.

Various network structures are known in the technique which more or less satisfactorily meet some of the above mentioned requirements .
For instance, in the system described in the American Patent No. 4,093,827 in the name of Thomson-CSF, a network is described equipped with a symmetrical time-division operating matrix which integrates the functions of signal memory, control memory, and interface towards PCM lines, serving channels organized into 8 groups.
More particularly the Fig. 11 of said patent illustrates a five-stage network (OE, A, B, C, OS) using switching elements implemented in the form of large-scale integrated circuits.
For said network the controls are performed by using a system with two hierarchical levels of which the higher (UC) directly controls the 1st and 5th stage and uses the lower control (MPP) to perform the control of the three intermediate stages (2nd, 3rd, and 4th stages).
The whole network is provided with a single unit UC and 8 MPP units.
In said American Patent the requirements referred to at points a), c), d), g), h), i), 1), m), are fully met; those referred to at e) are not satisfactorily met as the control structure is not adequately distributed.
The requirements referred to at b), f), on the contrary are not met in the above-mentioned patent, since the used symmetrical matrix is not directly controlled and not even controllable by a microprocessor of the commercially available kind nor is the diagnostic decentralized to the level of single switching module.
The symmetrical matrix is not directly controlled by a microprocessor since its interface towards the control unit is of the synchronous serial type and therefore said matrix is not flexible.
Structures capable of adequately meeting either all the above mentioned requirements or those left partially or totally unsolved by the above mentioned American Patent are still unknown in the technique.
These problems are solved by the present invention of a . modular structure of a PCM-switched network with distributed-control and distributed-diagnostic which being equipped with a local microprocessor control-unit is extremely flexible in the implementation of the typical switching functions; it is open to possible future technological developments of the integrated elements used; it achieves a considerable control distribution allowing the ratio between the processing power, installed under any equipment conditions, and the number of served channels to be optimized; being equipped with auxiliary circuits specialized in diagnostic it allows the operations of diagnostic, failure localization and reconfiguration to be expanded to the level of every building unit, covering also the failures concerning the connection between the single switching modules forming the network, without requiring the generation of an artificial test traffic of the connection and consequently, with negligible circuit overload, it achieves a quick and reliable identification of the single building unit in trouble, and the consequent system identification so that it might isolate the failing unit without any considerable prejudice to the service continuity.
According to this invention there is provided a modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages, belonging to a centralized-control exchange, consisting of modular connections basically comprising:

- integrated switching-matrices equipped with a microprocessor-compatible asynchronous control system;
- support circuits of an integrated type for diagnostic; and
- a microprocessor forming the lowest level of a three-level hierarchical control network;

- a plurality of peripheral connections, folded from the building standpoint, comprising the 1st and 5th time stages of said network;
- a plurality of central connecting units, folded from the building standpoint, comprising the 2nd and 4th time stages; and
- a plurality of unfolded central connecting units comprising the 3rd time stage of the same network;

An embodiment of this invention will now-be described, by may of example, with reference to the accompanying drawings wherein:

- Fig. 1 is the general scheme of the PCM-switching network comprising, as elementary modules, the connecting units (UCC, UCP) provided by the present invention;
- Fig. 2 is the block diagram of the peripheral connection unit, folded from the building standpoint, denoted by UC P1 in Fig. 1;
- Fig. 3 is the block diagram of the central connecting unit, folded from the building standpoint, denoted by UCC1 in Fig. 1;
- Fig. 4 is the block diagram of the unfolded central connecting unit denoted by UCC'1 in Fig. 1;
- Fig. 5 represents the hierarchical organization of the network controllers represented in Fig. 1.
- Fig. 6 shows an alternative scheme with respect to Figures 2, 3 and 4, considered together.

Fig. 1 represents by way of example, the structure of a switching network with 5 time stages with a capacity of 64,000 terminals.
The following concepts, as will be seen hereinafter, can also be applied to network structures, which differ in channel capacity and in number of stages, by any person skilled in the art.
ME1, ME2, ... , ME256 denote 256 PCM switching elements apt each to process 8 incoming PCM groups out of the 2048 groups (Fl, F8, F9, F16, ... F2041, F2048) connected to the switching network and coming from line units UL, not represented in the drawing.
These 256 switching elements, which are the first time stage (lT) of the above-mentioned network, are described in the Patent Application No. 67745 -A/80 filed on 13 May 1980 in the name of the Applicant.
They are PCM switching elements each of them comprising: means for converting the bits of incoming channels from series to parallel, a signal memory for storing the digital samples of the incoming channels, means for converting the bits of the outgoing channels from parallel to series, a control memory for storing the connections between incoming and outgoing channels, a control logic apt to receive and process data and commands coming from a control unit, and timing means, in which said control memory is able to store per each outgoing channel an additional bit ("busy" bit) sent by the control unit via the control logic, having the task of causing within the time slots of one or more outgoing channels, to which it is associated at an appropriate logic level, the inhibiting of said means for converting from parallel to series, in order to allow outgoing PCM groups of more switching elements to be connected in parallel, said inhibiting been achieved by replacing digital samples outgoing from said signal memory with words at constant logic level, said control logic being capable of receiving data and commands directly from an asynchronous control-unit on parallel data--bus and on control bus, and of processing data and commands sent by a control unit which is a microprocessor.
This switching element is made as an integrated circuit and can be associated to other similar switching elements to build up a higher capacity switching element.
References UC1, ..., UC8 denote 8 central units or switching planes, each consisting of 3 time stages (2T, 3T, 4T) using switching elements (MCE, MCC) of the same kind as those forming blocks ME1, ... , ME256, but apt to handle each 16 PCM groups instead of 8.
More particularly, as shown in detail for unit UC8, each central unit contains :

- 16 matrices MCE1, ... , MCE16 or 16 groups connected as in the Figure to matrices ME1, ... , ME256, such that the first group outgoing from the first switching element (ME1) of the lst stage (lT) is connected to the input No. 1 of central unit UC1, the second group outgoing from ME1 to the input No. 1 of central unit UC2, and so on up to the 8th group outgoing from ME1 which will be connected to the input No. 1 of central unit UC8. In this way it comes out that all the first groups outgoing from the 256 matrices ME1, ... , ME256are orderly connected to the 256 inputs of the central unit UC1; all the 2nd groups outgoing from the same 256 matrices are orderly connected to the 256 inputs of the central unit UC2, and so on till the 8th groups outgoing from the 256 matrices are orderly connected to the 256 inputs of the unit UC8.
- 16 matrices MCC1, ... , MCC16 each of 16 groups, analogous to the just examined matrices MCE1, ... , MCE16, connected at the input to the outputs of 16 matrices MCE1, ... , MCE16 as in the drawing, i. e. such that the first outgoing group of matrix MCE1 is connected to the input 1 of matrix MCC1, the second outgoing group to the 1st input of matrix MCC2, till the 16th outgoing group of matrix MCE1 is connected to the 1 st input of matrix MCC16. Analogously the 1st group outgoing from MCE2 is connected to the 2nd input of the matrix MCC1, till the 16th group of matrix MCE16 is connected to the input 16 of MCC16. In this way it comes out that the first groups outgoing from the 16th matrices MCE1, ..., MCE16 are orderly connected to the 16 inputs of the matrix MCCl, the second groups outgoing from the same matrices MCE1, ... , MCE16 are orderly connected to the 16 inputs of matrix MCC2, and so on till the l6th groups outgoing from said 16 matrices MCE1, ... , MCE16, are orderly connected to the 16 inputs of the matrix MCC16.
- 16 matrices MCU1, ... , MCU16 each composed of 16 groups, analogous to the ones above, connected at the input with the outputs of the 16 matrices MCC1, ... , MCC16 in an orderly way analogous to that seen for the just described connections between the 16 matrices MCE1, ... , MCE16 and the 16 matrices MCC1, ... , MCC16.
- 256 matrices MU1, ... , MU256, consisting of 8 groups of the same kind as those of matrices MEl, ... , ME256, forming the 5th time stage (5T), the inputs of which groups are connected to the outputs of the 8 central units UC1, ... , UC8 orderly and in a specular way with respect to the already described connections between the 25"6 matrices MEI, ... , ME256 of the first stage (1T) and the 8 central units UC1, ... , UC8.

As shown from Fig. 1, the central stage (3T), consisting of 128 matrices MCC (16 matrices per 8 UC units), is placed in symmetrical position with respect to the rest of the network.
Fig. 1 shows in dotted areas (denoted by UCP1, ... , UCP128, UCC1, ... , UCC16, UCC'1, ... , UCC'8) some sets of switching elements and of relative controllers which correspond to actual building modules (printed circuit boards) of the network, hereinafter referred to as peripheral connecting units (UCP) and central connecting units (UCC) folded from the building standpoint, and examined in detail in Fig. 2, 3 and 4. The fact of organising into an only building module UCP switching elements of the 1st and 5th time stage (IT, 5T) and into (another single module) matrices of the 2nd and 4th time stage (2T, 4T) each module being controlled by an only controller CTR1, allows the network shown in Fig. 1 to be used as a folded network.
This means the presence on the same connecting unit of both the incoming group and the corresponding outgoing group, which both result diagnosed by the same controller CTR1.
Moreover the folding from the building standpoint of connecting units UCP and UCC allows the connecting capacity of the single unit (1024 PCM channels) to be subdivided into two blocks of 512 channels used in the first (1T)andfifth (5T) stages (UCP) and in the second (2T) and fourth (4T) stages (UCC) respectively.
By this method it is obtained a modular growth step of overall network capacity, corresponding to 512 PCM channels rather than to 1024 channels, obtainable by the use of the single connecting unit in an unfolded way.
Moreover by this method, in case of failure of any one of said building units, the number of links run out of service can-be limited to 512.
The network modular-growth is achieved by progressively increasing at the periphery the number of peripheral connection units UCP, and inside each switching plane UC by progressively increasing the number of folded central connection units UCC.
As a consequence, an increment by a folded central connection unit UCC, on each of the 8 switching plans UC, corresponds to each increment equal to 8 connection units UCP.
The described interstage connection makes the network a complete-accessibility network. The advantages relating to blocking probabilities, growth modularity and degradation modalities presented by this type of structure are well known.
The symmetrical complete-accessibility network, so far examined, is nothing but the exchange switching network.
This network is to be controlled by a control network which, for diagnostic and reliability purposes ought to be decentralized.
As previously mentioned, in the patent in the name of Thomson-CSF, the controls are effected by using a two-hierarchical level system.
In our application, as shown also in Fig. 5, the switching network controller is further decentralized with respect to the centralized control unit CC of the exchange and is divided into three hierarchically-organized levels, denoted by CTR3, CTR2-CP, CTR1.
The portion which does not belong to such groups consists of controllers of 2nd hierarchical level (CTR2-CP) and of the controller CTR3 of the whole network; it is just a portion of the whole network.
CTR3 is the 3rd hierarchical level controller, i. e. the hierarchical highest one, and is generally obtained with a microprocessor technology of known type.
CTR3 receives from the control unit(CC), entrusted with telephone signalling processing, not shown in Fig. 1, the information on connections and disconnections to be established between the input and output of the whole network. For the connections. commands CTR3 searches for a connection path between the 1st stage (IT) and the 2nd (2T) and between the 4th stage (4T) and the 5th (5T) based on the busy state of the interstage connections and applying known algorithms for the minimization of the transit delay in said stages.
Once found a possible connection path, CTR3 entrusts the lower-level controllers (CTR2-CP) with the practical connection establishment.
The second-level controller (CTR2-CP) is also obtained by a microprocessor technology of known type. The tasks of the 2nd level controller are assigned as follows.
Each controller CTR2 is to control a unit UC (Fig. 1) of 8,000 channels.
CTR2 receives and checks the connect and disconnect messages sent by CTR3, by searching for a route, in case of connect command, or by disengaging the busy paths, in case of disconnect command. Connections between the second time stage (2T) and the third (3T) and between this one (3T) and the fourth (4T) are effected based on the busy state of interstage connections and by applying a method for transit delay minimization in said stages.
Every 16 1st-level controllers CTR1, there is one peripheral unit controller CP.
Every CP receives and checks the messages relating to connect and disconnect messages received from CTR3 and forwards to the 1st-level controller (CTR1) the connect and disconnect commands.
The Ist-level controller CTR1, the hierarchically-lowest, is also obtained with a well-known microprocessor technology. CTR1 controls the central and peripheral connecting units, and executes the following operations:

- reception and check of connect and disconnect commands received from CTR2 and CP;
- connect and disconnect commands to the switching elements of Fig. 1 composing the single switching units.

How the connect and disconnect commands sent by the central exchange control system (CC) are executed has been examined so far.
The problem of failure localization and diagnosis is solved by equipping all the switching units with their own self-diagnostic system, carried out by the 1st-level controllers CTR1 and by suitable support circuits (CDT, RTB), which will be examined hereinafter, built in each connecting unit; thus the unit becomes nearly autonomous (self-sufficient) with regard to failure detection and localization. For the portion of the network (CTRZ-CP, CTR3) not implemented by the connection unit the failure diagnostic and reconfiguration can be obtained by known methods.
Fig. 2 shows the peripheral connection unit denoted byUCPl in Fig. 1. In it references IT and 5T denote the same time stages as shown in Fig. 1.
MEI, ME2, MU1, MU2, CTR1 denote the same blocks as shown in Fig. 1.
In Fig. 2 CDTE1, CDTE5 and CDTU1 CDTU5 denote 4 sampling circuits for transfer diagnosis, the first two being placed at the incoming side and the other two at the outgoing side of the peripheral connection unit UCP1.
Said sampling circuits consist each of an integrated circuit, on a large integration scale, performing the extracting function of a _'bit octet relating to a channel, with a pre-defined frame delay, from a PCM group inside the set of 16 groups entering ME1, ME2 and MU1, MU2. The octet is stored and can be presented on the data bus towards the lst-level controller CTR1.
Circuits such as CDTE1 CDTE5 are described in the Patent Application No. 67259-A/80 filed on the 20th February 1980 in the name of the Applicant.
References RTBE1, RTBE5 and RTBUl, RTBU5 denotefour simultaneous bidirectional transceivers; the first two placed on the incoming side of the first (IT) and fifth (5T) stages respectively and the other two on the outgoing side respectively of the same stages.
Circuits of type RTBE and RTBU are described in the Patent Application No. 68914-A/79 filed on 14, October 1979 in the name of the Applicant.
The bidirectionality of the transceivers (RTBE, RTBU) allows the same physical way to be used for both transmission directions.
These transceivers are used here to perform the diagnostic of connection continuity, by back-transmitting the signal which arrives at the circuit RTB placed at the far end of a line and thus allowing the comparison between sent signal and back signal (echo check) without requiring further interstage connections.
RTBE_UL and RTBU_UL denote transceivers analogous to the preceding ones inserted on the incoming and on the outgoing side of the line units (UL) placed at the periphery with respect to the connection network.
The first-level controller CTR1, as a rule, consists of a microprocessor which controls, through the bidirectional bus bd, matrices ME1, ME2, MU1, MU2 and the samplers CDTE1, CDTE5 and CDTU1, CDTU5. Besides controller CTR1 through the bidirectional connection bc, can dialog with the higher-level controller CP.
The operation of the circuit whose block diagram is depicted in Fig. 2 is the following: 16 incoming groups coming from line units (UL) which, through the transceiver RTBE1, enter ME1, ME2 as well as the incoming sampling circuit CDTE1, are connected on the incoming side of the peripheral connecting unit (UCP1).
Matrices MEI, ME2, upon the switching command received from CTR1 through bus bd, perform the suitable, switching operations between the channels of the incoming and outgoing groups.
The 16 groups outgoing from ME1, ME2, belonging to the first time stage (IT), are respectively transmitted through the outgoing transceiver RTBU1 to the incoming transceivers RTBE2 of the subsequent time stage (2T).
Each group received by an incoming transceiver RTBE2 of the second stage (2T) is re-transmitted to the outgoing transceiver RTB1 of the previous stage (IT), by which it has been generated, and then sent to the outgoing sampler CDTU1.
On the outgoing side of UCP1 there are connected 16 groups coming from transceivers RTBU4, relating to the 4th stage switching unit (4T) which through transceiver RTBE5, matrices MU1, MU2, samplers CDTE5, CDTU5, are processed analogously to what describedfor the incoming side of UCP1 in order to obtain 16 outgoing groups sent to transceivers RTBE_UL of line units (UL).
More particularly Fig. 3 shows the folded connection unit UCC1 of central unit UC8.
In that Figure MCE1, MCU1, CTR1 denote the same blocks shown in Fig. particularly the matrices MCE1 and MCU1 are obtained by interconnecting under matrix format, four switching elements EC1, EC2, EC3, EC4, of the ME1 kind made of 256 channels so as to obtain a single non-blocking switching matrix capable of connecting 512 channels.
This structural organization and its operation as well are also described in the Patent Application No. 67745-A/80 previously cited with reference to switching elements MEl... ME256 of Fig. 1; more particularly that structural organization is shown in Fig. 3 of the mentioned Patent Application.
Controller CTR1, among the other tasks which will be examined, hereinafter, ensures the command of the four switching elements contained in MCE1 and MCU1.
Blocks denoted by RTBE, RTBU, CDTE, CDTU are analogous to those shown in Fig. 2 and are connected in the same way.
Connection groups denoted by a₁, a₂, a3, a4 are exactly the same as those of Fig. 1.
Bus bd and connection be are identical to those shown in Fig. 2.
In particular, Fig. 4 shows the unfolded connecting unit UCC'l of central unit UC8.
In that Figure MCC1, MCC2, CTR1 and c₁, c₂ denote the same blocks and the same groups of wires shown in Fig.l; RTBE, RTBU, CDTE, CDTU denote blocks perfectly analogous to those shown in Fig.3.
With the organization of the switching elements MCE1, MCU1 MCC1, MCC2, given by way of example in Figures 3 and 4, double-capacity matrices, if available due to technological progress could be used without altering the structure of the whole network shown inFig.l.
With reference to Figures 2, 3, 4 it is worth noting that in case the retransmission of the group received by an incoming transceiver RTBE is effected on the same physical path where said group has been received, non-detection of a failure due to an interruption of the physical path itself (open circuit) may occur.
In this case, in fact, transceiver RTBU placed upstream the incoming transceiver RTBE would however receive because of reflection on the cable (open circuit) a false echo which might mislead the transfer diagnosis sampler.
A simple and easy way to overcome this disadvantage with the just described network resides in the use of the second physical way, provided together with connections among the various connecting units for the back transmission of the diagnostic group.
In fact, as shown in Fig. 1, for example matrix MCE of UCC1 of the 2nd stage is connected through only two physical paths (a_l) to matrices ME1, ME2 of UCP1; analogously two physical paths (a4) are provided for the connection between MCU1 of UCC1 and matrices MU1, MU2 of UCP1.
The same process applies to the connection paths among the matrices of all the folded central connecting units UCC and those of unfolded connecting units UCC'.
The capacity values as to the number of PCM channels processed by the single connecting units have been pre-determined so as to allow the implementation with the present technology of the connecting unit by a single replaceable structural element (board including connectors).
More particularly said capacity values have been determined on the basis of the properties

- of the used switching elements (ME, MCE, MCC, MCU, MU);
- of the support circuits designed for the diagnostic (CDT, RTB);
- of commercial microprocessors (CTR1), and
- of relative peripheral circuits taking also into account the specifications for the standard board size, such as for instance the "Doppio Europa" in agreement with the DIN specifications.

The utilization of network structures which make intensive and prevalent use of replaceable structural elements of the same kind, like those here examined, is advantageous from the manufacture, the stocking and maintenance standpoints.
Fig. 5 is a schematic representation of the hierarchical structure of central exchange control system (CC), of the 3rd level controller CTR3, of the 2nd level controllers CTR2-CP and finally of the 1st level controllers CTR1.
Numbers 1, 8, 9, 24, 16, 113, 114, 128 written near the CTR1 denote the number of the controllers CTR1 which are needed. Thus for instance for each UC (in the Figure UC1, ... , UC8) there are provided 24 CTR1; for each CP there are provided 16 CTRl. The arrows outgoing from CTR1 denote the switching units (UCC1, ..., UCC16, UCC'l,..., UCC'8, UCP..., UCP16, ... , UCP113, ... , UCP128) served by said controllers.
The diagnosis process is performed in a distributed way on connecting units (UCP, UCC, UCC') forming the network.
Since this process repeats analogously for the three described kinds of connecting units, for the sake of brevity only the diagnostic process relating to the folded central connection unit (UCC1 of Fig. 3) will be described hereinafter by way of example.
Through bidirectional connection be the 2nd-level control- ler CTR2 dependent on the routing search effected on UC8, sends to the folded central connection unit UCC1 an interconnect command between an incoming channel and an outgoing one; the latter may correspond to the second stage (MCEl) or to the 4th stage (MCU1).
CTR1 is capable of recognizing whether the "connect command" relates to MCE1 or to MCUI; by way of example let us suppose that MCE1 is involved and that the command is of connecting the channel a of the incoming group 1 with the channel b of the outgoing group 16.
CTRl, depending on the order number of the (incoming outgoing) group involved in the switching,selects the switching unit, namely EC1, on which the connection interesting the channel a of input 1 and channel b of output 8 of EC 1 is to be established.
CTR1 will have previously disabled channel b of the output 8 of EC2, which has its outputs connected in parallel (AND-wired) with the corresponding outputs of EC1. After that the connection is estabilish- ed.
Controller CTR1 can effect, according to intervention modalities which will be defined case by case, the diagnostic according to the procedure which will be hereinafter described, on any connection which is either already established or still in progress.
With reference to the chosen example, the diagnostic is performed by extracting an octet of the channel a of the input group 1 and by comparing it with the same octet of the channel b of outgoing group 16, depending on the route on the physical path a₂ of the connection towards the subsequent stage and by taking into account the delay relating to the frame unit introduced by the transit of the connection in MCE1.
In this way, a diagnosis is operated not only on matrix MCE1, but also on its connection with other matrices belonging to different stages.
By applying this method to all the network stages the "tangency" is obtained among the diagnostics of adjacent switching units involved in a complete connection between the network input and output, without leaving portions of the network undiagnosed.
More particularly, CTR1 calculates said transit delay, it sends the octets extracting commands to CDTE and CDTU respectively, and finally compares said commands between them.
If the comparison result is positive, CTR1 continues its control operation according to the modalities envisaged case by case; if the result is negative it means that a malfunctioning has occurred which may concern: matrix MCE1; transceiver RTBU of UCC1, or that of kind RTBE, connected to RTBU and belonging to the following stage; all the interconnections interested in the above-mentioned blocks.
The malfunctioning may also have interested the peripheral units of CTR1 and more particularly the samplers CDTE, CDTU and their respective connections.
The detection of possible troubles in the processing unit (microprocessor, memories) of CTRl is actuated by self-diagnosis methods of a known type (parity-checks, software-traps).
Also the testing of the good functioning of connection be among CTR2 and CTR1 is performed according to well-known methods, such as.the message parity check (check-sum).
Finally, the malfunctioning may however be localized inside UCC1, or may concern the transceiver RTBE of the subsequent stage and relative connection.
It is up to the higher-level controller CTR2, depending on the malfunctioning signallings it receives through the various CTR1 to proceed according to modalities defined while designing the apparatus, to isolate the unit considered in trouble and to effect the subsequent network reconfiguration; the latter may be effected by using whatever method known in the technique.
The diagnostic of the higher hierarchical portion of the control-tree branches (CTR1-CP, CTR3) is effected according to known methodologies.
The just described diagnostic process allows different methods for testing the connections. For instance it is possible to test the connection immediately after its establishment, to cyclically scanall or parts of the connections already established, or to effect both tests by assigning priority criteria which will be defined while designingthe system.
As a rule, the just established connection ought to be assigned maximum priority.
Besides, said diagnostic process is compatible with known global processes of network supervision, carried out by a supervising device which may be present in the exchange.
From what previously expaunded it is evident that the tasks of the 1st level controller CTRI are according to a priority order: execution of connect and disconnect commands coming from CTR2, diag- osis of the established connection, self-diagnosis; finally of course CTR1 is to send possible alarm signallings to CTR2 through connection be.
A different way of performing the interstage connection, exploiting the folding network and the bidirectional transceivers (RTB), already examined, is the one represented in Fig. 6.
The blocks of Fig. 6 are the same as those already examined in Figs. 2, 3 and 4.
Reference m₁ (Fig. 6) denotes the bidirectional support physical path for the PCM group connecting the switching matrixMEl of the 1st stage (IT) with the matrix MCE1 of the second stage (2T) as well as for the PCM group which connects the matrix MCU1 of the 4th stage (4T) with the matrix MU of the 5th time stage (5T) of the network.
Reference m₂ denotes the bidirectional support physical path for the PCM group connecting the switching matrix MCE1 of the second (2T) time stage with matrix MCC1 of the third (3T) stage and for PCM group connecting the latter matrix with the matrix MCU1 of the fourth (4T) time stage.
Reference m₃ denotes the bidirectional support physical path for the PCM group of diagnosis (echo check) relative to the connection between the second (2T) and the third (3T) stage and for the group PCM of diagnosis relative to the connection between the third (3T) and 4th (4T) time stage.
Reference m₄ denotes a bidirectional support physical path for the PCM group of diagnosis relative to the connection between the 1st (1T) and the2nd (2T) time stage and for the PCM group of diagnosis relative to the connection between the 4th (4T) and 5th (5T) time stage.
The peripheral connection unit UCP1, already described in Fig. 2, is shown in Fig. 6, modified in the connections relating to connections between RTBU1 and CDTU1 as well as in those between RTBE5 and CDTE5.
More particularly, the variation is that the sampler CDTU1, relative to matrix MEI, is connected to the transceiver output RTBE5, and not as in Fig. 2, to the output of transceiver RTBU1; and more particularly is that the input of matrix MUl (Fig. 6) is connected to the output of transceiver RTBU1 and not, as in Fig. 2, to the output of RTBE5.
Central connecting units UCC1 and UCC'1, already described in Figures 3 and 4, are represented in Fig. 6 with changes in connections concerning the connections between RTBU and CDTU as well as those between RTBE and CDTE.
In particular, the variation resides in that the sampler CDTU1, relating to matrix ME1, is connected to the output of the transceiver RTBE5 and not as in Fig. 2, to the output of transceiver RTBU1; and in addition in the fact that the input of matrix MU1 (Fig. 6) is connected to the output of transceiver RTB1, and not, as in Fig. 2, to the output of RTBE5.
The central connection units UCC1 and UCC'1, already described in Figures 3 and 4, are shown in Fig. 6 with changes in the connections relating to the links between RTBU and CDTU and those between RTBE and CDTE.
More particularly the variation for the folded central unit UCC1 of Fig. 3 is that sampler CDTU, relative to matrix MCE1, is referred to as CDTU2 in Fig. 6 and now is connected with its inputs to the incoming transceiver RTBE4 (Fig. 6), placed at the side of MCU1; the CDTE4 of the 4th (4T) stage is now connected to the transceiver outputs RTBU2, placed at the MCE1 side.
In addition, the MCUI output is connected to the input of transceiver RTBE2 and the input of matrix MCE1 is connected to the input of transceiver RTBU4.
An analogous modification takes place for the unfolded central unit UCC'1 (Fig. 4) that is why MCC1 outputs are connected to RTBE3 inputs and MCC1 input, instead of being connected as in Fig. 4 to RTBE, is connected to the RTBU3 input.
Figure 6 outlines how the physical path m₂ is used for instance as a support for the connection between the output of the 2nd stage matrix MCE and the input of the 3rd stage matrix MCC1 and at the same time for the connection between the output of matrix MCCI and the input of the 4th-stage matrix MCU1.
As to the diagnostic function of the physical connections allowing the attainment of a diagnosis "tangency" among the various network stages, a second bidirectional physical support m₃ is to be used, which in the connecting scheme depicted in Figures 3 and 4 was the connection physical path c₂ between the 3rd and 4th time stages.
The advantages derived by carrying out the interstage connections according to the block diagram of Fig. 6 are:

- complete separation of the paths for the groups of interstage connections, and their diagnostic, so that it is no longer necessary to condition the connections and their diagnostic as previously described by effecting them on different physical paths;
- halving of the number of cables used in the network when interstage connection diagnosis is not required. In this case there are only bidirectional connection paths m_l, and m_Z and there is no longer the already examined diagnostic tangency. It will be sufficient to establish the connections represented in dotted lines in Fig. 6 by p₁, P₂, P3 and P4 placed at the outputs of the switching units, and more particularly of MCC1, MCU1, MCE1, ME1 respectively.

Said connections p₁, ... , p4 by directly connecting the matrix outputs with the inputs of "transfer diagnosis samplers" (CDTU) for the transfer of outputs lead the process of diagnosis of the connection units unchanged.
It is clear that the above considerations apply, by means of simple adjustments which can be performed by any one skilled in the art, to networks with a different channel capacity or with a higher or lower number of time stages.
For instance if the number of stages is to remain unchanged and the whole network capacity is to be doubled, by bringing it from _'65,536 channels to 131,072 channels, it is sufficient to double planes UC (Fig. 1) bringing them from 8 to 16 and substituting the peripheral connection units UCP by a· double number (256) of central connection units UCC, folded from the building standpoint..
Resorting to units UCC instead of resorting to UCP, is due to the necessity to maintain the full accessibility to the 16 planes of UC.
To obtain a network with the same capacity as 65, 536 channels, but with a number of time stages equal to 7, it would be sufficient to use through the whole network only connection units of the UCP kind.
More particularly the present invention may cover network structures whose capacity ranges from few hundred PCM channels to various hundreds of thousands of channels with a number of time stages comprised between one and seven time stages.
Modifications and variations are possible without going out of the scope of the invention.

Claims

1. Modular structure of PCM-switched distributed control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, characterized in that it consists of modular connections basically comprising:

- integrated switching-matrices (M, MC) equipped with a microprocessor-compatible asynchronous control system;

- support circuits of an integrated type (RTB, CDT) for a diagnostic; and a microprocessor (CTR1) forming the lowestlevel of a three-level hierarchical control network; said modular connecting units, in case of five time stages being functionally diversified in

- a plurality of peripheral connections folded from the building standpoint (UCP), comprising the 1 st (IT) and 5th (5T) time stages of said network;

- a plurality of central connecting units folded from the building standpoint (UCC), comprising the 2nd (2T) and 4th (4T) time stages; and

- a plurality. of unfolded central connecting units comprising the 3rd time stage (3T) of the same network.

2. Modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to claim 1, characterized in that each of said peripheral connection units (UCP), of said folded central units (UCC) and of said unfolded central units (UCC') is physically contained in a single standardized plate, which can be extracted and replaced, said plates being the modular structural element of said network structure.

3. Modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to claim 1 and 2 characterized in that the control at the level of single modular structural element is effected by the microprocessor (CTR1) present in each one of the central connection (UCC, UCC') and peripheral (UCP) units; said microprocessor (CTR1) sending to the switching units (ME, MU, MCE, MCU, MCC), the connect and disconnect commands it receives from the controller with an immediately higher level (CTR2).

4. Modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to claim 1, 2 characterized in that each of said peripheral connection units (UCP) is connected to a central folded connection unit (UCC) via two output physical paths forming the connection between the outputs of the 1 st (IT) time stage and the inputs of the second stage (2T) and via two input physical paths, forming the connection between the outputs of the 4th (4T) time stage and the inputs of the 5th (5T) stage.

5. Modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to claims 1, 2, characterized in that each of said folded central units (UCC) is connected to a unit of said unfolded central units (UCC') via two physical paths forming the connection between the outputs of the second (2T) time stage and the other inputs of the 3rd (3T) stage and via two input physical paths forming the connection between the outputs of the 3rd (3T) time stage and the inputs of the 4th (4T) stage.

6. Modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to claims 1, 2, 3, 4, characterized in that the diagnostic at level of each modular switching unit is performed in conjunction by said microprocessor (CTRI), belonging to each plate, and by said support circuits, consisting of sampling circuits (CDTE, CDTU) for the diagnosis of the transfer of PCM signallings within the time stage it belongs to and within the interconnection between the 4th time stage and the next, and consisting also of simultaneous bidirectional transceivers (RTBE, RTBU) used to diagnose a connection continuity, each transceiver, placed at the far end of a connection path (line), back-transmitting the signal it receives so as to allow the comparison (echo check) between the back signal and the sent one, said signals being forwarded on different physical paths.

₇. Modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to claim 6 characterized in that to obtain said diagnostic within a plate said sampling circuits (CDTE, CDTU) extract a bit octet of the incoming channel (a) which is to be switched in the time stage and compare it with the same octet extracted on the relative outgoing switched channel (b), taking into account the delay introduced by the transfer within the same time stage, said comparison, said transit delay computation, and octet extract commands being managed by said microprocessor (CTRl).

₈. Modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to claim 1, characterized in that said sampling circuits (CDTE, CDTU) and said simultaneous bidirectional transceivers (RTBE, RTBU) cooperate to diagnose both the same structural unit and the connections between said structural unit and the contiguous ones so as to obtain the "tangency" among the diagnostics of the contiguous structural units interested in a complete connection between the network input and output; said cooperation being implemented by inserting between the switching matrices (MCC, MCE, MCU, ME, MU) and the relative samplers (CDTE, CDTU) belonging to a structural unit a number of bidirectional transceivers (RTBE, RTBU) belonging to contiguous structural units.

9. Modular structure of PCM switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to the previous claims characterized in that the structural element performing as peripheral connection unit (UCP1) consists of:

- 4 simultaneous bidirectional transceivers (RTBE1, RTBE5, RTBU1, RTBU5) used to diagnose the continuity of the connection relating to 16 PCM groups, on the basis of the "echo check" criterion;

- a lower-level control (CTR1);

- 4 integrated switching matrices (MEI, ME2, MU1, MU2), capable of effecting, on the basis of the switch commands coming from said controller (CTRI), the suitable switching operations between the channels of the incoming groups and those of the outgoing groups;

- 4 sampling circuits (CDTE1, CDTE5, CDTU1, CDTU5) for the transfer of the bit octet diagnosis within the time stage they belong to.

10. Modular structure of PCM. switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange, according to the previous claims characterized in that the structural element, performing as folded (UCC1) and unfolded (UCC'1) central connection unit, consists of:

- 4 simultaneous bidirectional transceivers (RTBE, RTBU) capable of diagnosing the continuity of the incoming and outgoing connections on the basis of the "echo check" criterion;

- 2 non-blocking connection matrices (MCEI, MCU1; MCC1, MCC2) capable of connecting 512 channels;

- a lower-level controller (CTR1);

- 4 sampling circuits (CDTE, CDTU) for the bit-octet transfer diagnosis within the time stage they belong to.

11. Modular structure of PCM-switched distributed-control and distributed-diagnostic network with a plurality of time stages belonging to a centralized-control exchange according to claims 1, 2, characterized in that interstage connections between each peripheral connection unit (UCP1) and each folded central connection unit (UCC1) are implemented by a first pair of bidirectional transceivers (RTBU1, RTBE2) which are connected by a first bidirectional connection (m_l) in order to connect the input of the first (1 T) time stage with the input matrix of the second (2T) and at the same time to connect the output of the matrix of the 4th (4T) time stage with the input of the matrix of the 5th (5T) stage, the diagnostic of said interstage connections being effected through a second pair of bidirectional transceivers (RTBU4, RTBE5) which are connected by a second bidirectional connection (m₄) and in that the interstage connections between each unfolded central unit (UCC1) and each unfolded central connection unit (UCC'1) are obtained by means of a third pair of bidirectional transceivers (RTBU2, RTBE3) which are interconnected by a third bidirectional connection (m_Z) to connect the output of the second (2T) time stage with the input of the third (3T) stage and at the same time the output of the third (3T) time stage with the input of the 4th (4T) stage, the diagnostic of said interstage connections being effected through a fourth pair of bidirectional transceivers (RTBU3, RTBE4) which are connected by a fourth bidirectional connection (m3).